Abstract. Respiration in tree stems is an important component of forest carbon balance. The rate of CO 2 efflux from the stem has often been assumed to be a measure of stem respiration. However, recent work in temperate forests has demonstrated that stem CO 2 efflux can either overestimate or underestimate respiration rate because of emission or removal of CO 2 by transport in xylem water. Here, we studied gas exchange from stems of tropical forest trees using a new approach to better understand respiration in an ecosystem that plays a key role in the global carbon cycle. Our main questions were (1) is internal CO 2 transport important in tropical trees, and, if so, (2) does this transport result in net release of CO 2 respired in the roots at the stem, or does it cause the opposite effect of net removal of stem-respired CO 2 ? To answer these questions, we measured the ratio of stem CO 2 efflux to O 2 influx. This ratio, defined here as apparent respiratory quotient (ARQ), is expected to equal 1.0 if carbohydrates are the substrate for respiration, and the net transport of CO 2 in the xylem water is negligible. Using a stem chamber approach to quantifying ARQ, we found values of 0.66 ± 0.18. These low ARQ values indicate that a large portion of respired CO 2 (∼ 35 %) is not emitted locally, and is probably transported upward in the stem. ARQ values of 0.21 ± 0.10 were found for the steady-state gas concentration within the stem, sampled by in-stem equilibration probes. These lower values may result from the proximity to the xylem water stream. In contrast, we found ARQ values of 1.00 ± 0.13 for soil respiration. Our results indicate the existence of a considerable internal flux of CO 2 in the stems of tropical trees. If the transported CO 2 is used in the canopy as a substrate for photosynthesis, it could account for up to 10 % of the C fixed by the tree, and perhaps serve as a mechanism that buffers the response of the tree to changing CO 2 levels. Our results also indicate, in agreement with previous work, that the widely used CO 2 efflux approach for determining stem respiration is unreliable. We demonstrate here a field applicable approach for measuring the O 2 uptake rate, which we suggest to be a more appropriate method to estimate stem respiration rates.